Email warmup in 2026: what actually moves inbox placement

Warmup is the most talked-about and least understood part of cold email infrastructure. Every tool offers it, every guide says to do it, and yet plenty of teams warm mailboxes for three weeks, switch on a campaign and watch placement slide within a fortnight anyway. This post covers what warmup actually does, what it cannot do, and how to run it so the effort shows up in the inbox rather than in a dashboard.

What warmup is really for

Mailbox providers decide where to put a message based on the reputation of the sending domain, the sending IP and the mailbox itself, and on how recipients have responded to mail from those three things in the past. A brand-new mailbox on a brand-new domain has no history. Providers treat no history as risk, and risk means the spam folder or throttled delivery.

Warmup manufactures history. A warming mailbox exchanges messages with a network of other mailboxes that open them, reply to them, mark them as important and rescue them from spam. To the provider, that looks like a normal person starting to use a new account. Volume ramps gradually so the account never looks like it appeared out of nowhere and started blasting.

That is the whole mechanism. It is useful, and skipping it is a mistake, but it is worth being clear about what the mechanism can and cannot achieve.

What warmup cannot fix

A bad list. If a large share of your recipients do not exist, warmup history evaporates fast. Bounces are the strongest negative signal you can send, and a warmed mailbox that hard-bounces on its first campaign is right back where it started. Verify every list before it touches a mailbox.

Broken authentication. SPF, DKIM and DMARC are table stakes, and the major providers have tightened enforcement year after year. Warming a domain with missing or misaligned DKIM is warming a domain the provider does not trust. Set authentication up before the first warmup message, not after the first campaign.

Sending too much, too soon. Warmup earns you a reputation budget. Blowing through it on day one of a campaign, by sending far more cold mail than the mailbox ever sent during warmup, spends the budget in a day. Volume after warmup should step up, not jump.

Copy that gets reported. Recipients marking your mail as spam is a stronger signal than any positive warmup interaction. If your copy is generating complaints, more warmup will not save it.

How long warmup takes

There is no single answer, because the goal is a stable reputation rather than a calendar date. As a planning figure, allow two to four weeks before a fresh mailbox carries meaningful cold volume, with the ramp roughly following this shape:

Google and Microsoft mailboxes tend to behave differently through this curve, and independent SMTP infrastructure behaves differently again. This is one of the arguments for running a mix of all three rather than betting the whole program on one provider's mood.

Warmup is a maintenance schedule, not a phase

The single biggest change in how experienced senders think about warmup is that it no longer ends. A mailbox that stops receiving positive engagement and only sends cold mail slowly becomes a mailbox whose entire history is cold mail. Reputation drifts, placement follows.

Keeping a background level of warmup traffic running alongside live campaigns does two things. It keeps a steady stream of positive signals attached to the mailbox, and it gives you a canary: if warmup messages from a mailbox start landing in spam, that mailbox's cold mail is landing there too, and you have found out before the campaign stats told you.

Measuring placement instead of guessing

Open rates are a poor proxy for placement. Image blocking, privacy features that pre-fetch pixels and corporate gateways that scan links all distort them. If you want to know where a send landed, you need signals that are harder to fake:

1. Seed-list placement tests. Send to a panel of mailboxes you control across providers and check where each message arrives.
2. Bounce and deferral codes. Providers tell you a lot in their SMTP responses if you read them per mailbox rather than in aggregate.
3. Blocklist monitoring. A domain or IP on a major list should be pulled from rotation the hour it appears, not when the weekly report comes out.
4. Reply and complaint ratios per mailbox. Trend these individually. A pool average hides the two mailboxes dragging everyone down.

Per-mailbox monitoring is what turns warmup from a ritual into a feedback loop. When you can see that a mailbox has drifted, you can slow it down, re-warm it or retire it before it spends the reputation of the domain it shares with others.

Rotation and cycling: the part most guides skip

Warmup keeps mailboxes healthy. Rotation keeps any single mailbox from carrying too much of the load. Cycling replaces mailboxes that have run their course with fresh, pre-warmed ones. All three together are what produces a stable inbox rate over months rather than a good first fortnight.

Doing this by hand means a recurring chore: build new domains and mailboxes ahead of need, warm them while the current set is live, watch the per-mailbox numbers, swap under-performers out and never let the buffer shrink. It is entirely possible. It is also the first thing that gets dropped when the team is busy.

Putting it together

If you run your own infrastructure, the checklist is: authenticate every domain before warming, warm for weeks rather than days, step volume up gradually, keep warmup running for the life of the mailbox, verify lists, and measure placement per mailbox so you can act on drift early.

If you would rather that all happened without a calendar reminder, that is what Sender Pools does. Every mailbox in your pool is warmed before it sends cold mail, kept warm after, monitored individually and cycled out when its numbers slip. It is how we hold a 97.4% inbox placement rate across more than a million emails a day. The pricing page explains what that costs, and what it does not.